So, what else is going on?

In the residential sector, growth stalled after 2005 and it wasn’t until 2017 that it returned. 

The chart below shows the average kWh consumed per occupied household from 1991 to 2025. It has fallen since 2005, and is still falling today.

I am not aware of any definitive study on why household energy use has fallen, but there are probably a range of factors at play.

The climate is warming, which means less energy is required to heat homes.  Furthermore, heat pumps are now accepted as a reliable and cost-effective heating appliance, and they are many times more efficient than direct electricity heating, e.g. convective and radiant heaters.

When was the last time you bought an old incandescent light bulb?  LED lights are now everywhere, and virtually the only type of light used in new buildings. An old-fashioned incandescent light bulb is basically just a heater that emits light, whereas LED lights emit light and very little heat.

Appliances have become more efficient since 2002 when the Energy Efficiency (Energy Using Products) Regulations came into effect and the Energy Efficiency and Conservation Authority started mandatory appliance testing programs.

The price of electricity rose in real terms from 2002 to 2024, which probably caused some consumers in the North Is, where there is access to piped natural gas, to switch their heating to gas.

The average size of new dwellings has fallen, and changes to the NZ Building Code progressively improved the energy efficiency of new builds. It is now standard practice to install insulation when renovations are undertaken on older dwellings.

Demand in the commercial and ag-fish-forestry sectors bucked the trend by continuing to rise after 2006, hitting peaks in 2019 and 2024, respectively, driven by the growth of dairy production and irrigation, in particular. The demand at the Ashburton supply point in Canterbury shows a very marked peak in the spring and summer when irrigation is usually at its peak over 180 MW, then a large drop into winter as milk production falls to sit under 70 MW.

Commercial demand is the sector we understand the least, and it also remained almost constant since 2019. It currently accounts for 23% of annual demand, and is neither growing nor shrinking right now. No doubt energy efficiency has played a role in office buildings in particular, i.e. building code, heat pumps and LED lights, along with the impact of rising temperatures.

But what else is happening in this sector? Are we seeing, for example, the impact of online shopping reducing the number and scope of retail operations, leading to lower demand for electricity?

Commercial demand correlated well with real national GDP until about 2012, after which the economy (represented by GDP) continued to grow but commercial demand remained flat. From 2015 to 2025, for example, the total number of businesses in NZ grew from 489,000 to 599,000 (+22.2%) while commercial demand went from 9,455 GWh to 9,333 GWh (-1.3%) which suggests to me there is more going on than just an increase in the efficient use of electricity.

Commercial demand reached a peak of 9,356 GWh in 2019 but then fell to under 9,200 GWh in 2020 (-3.7%), after which it grew again to 9,448 GWh in 2023 and then fell back to 9,332 GWh in 2025. Obviously, covid had a huge impact on this sector as we all worked from home, or didn’t work at all, in 2020, but subsequently people are working from home more, which means that workplaces can downsize or ‘no-size’ their office space. 

Online shopping had to increase in 2020 during lockdowns, and has remained permanently higher than the pre-covid days, which leads to less or smaller physical shops.

The trend toward lower demand growth seen in NZ since 2006 is not peculiar to NZ, with countries like the UK, the US and Australia also showing a change in the rate of growth. In the UK, for example, demand fell significantly after 2006. So NZ is not alone in experiencing de-industrialisation, increasing energy efficiency and changes in the way that we live, work and shop.

Looking Ahead

Demand growth more or less disappeared after 2006 but almost everyone, including Energy Link, is forecasting demand growth to return. Furthermore, as we saw in the second article in the series, there is a veritable renaissance going on in new renewable generation, which must be partly or wholly due to an expectation of significant growth in demand and a pressing need for more generation to serve that demand.

But how likely is it that demand growth will return?

First, I have to say that forecasting demand is difficult. Government has got it very wrong in the past and the market got it very wrong in the first few years after 2006. 

Energy Link’s forecasts are made up of five demand scenarios which cover a wide range. Yes, the lowest and highest scenarios have a low probability assigned to them, but they are still possible.

On the low side, the wood, paper and pulp processing sector could continue to shrink if the long-term trend continues. Other sectors could shrink if they cannot find a cost-effective replacement for natural gas for producing process heat, and businesses close as a result. Rising temperatures will keep downward pressure on demand in buildings, until such time that large areas of NZ require cooling in the warmer months, which is still years away.

On the high side, the fourth potline at the Tiwai Pt smelter will restart sometime before the end of 2028, Fonterra has major electrification projects underway, and a hyperscale data centre could be built in Southland to eventually match or even surpass the annual consumption of the smelter, currently around 12% of total annual consumption.

But despite the stagnation of demand since 2006, all the signs are that demand growth will return. The chart below shows the demand for each island and for NZ relative to the same time last year, each excluding Tiwai Pt demand, and Tiwai is shown separately.

Demand for the year to 31 August is just over 300 GWh higher than it was last year at the same time, although 250 GWh of this is due entirely to the Tiwai Pt smelter returning to its normal demand level after having to turn down in 2024 and 2025 to help keep lake levels up during the dry period of 2024. 

Electric vehicles (EVs) have the potential to add 5,000 GWh per annum or more to demand, in the long term but so far, EV sales have a checkered history.

The chart below shows the monthly registrations of EV and plug-in hybrids since 2019 (the great majority are pure battery EVs). They took off when the so-called ‘ute tax’ was introduced and you could get up to $7,500 rebate on a new EV costing up to $80,000.

This saw EV registrations rise to an average of 15% in 2023, after which they plummeted to less than 5% when the subsidy ended.

But recently, the market is flooded with cheap and solid EV offerings from China which, along with a sharp rise in the price of petrol and diesel when ‘you-know-who’ started a war with Iran, has caused registrations to hit an average of 17% in the last five months.

There are two EVs in our household and we love the way they drive, their reliability and their low, low running costs. Given their falling as-new prices, increasing availability of used EVs, and the ‘network effect’ caused by seeing your friends and neighbours with EVs, it is inevitable that demand from EVs will grow. Sales of new technology are known to follow a flattened ‘S curve’, with sales starting slowly then ramping up over time, before starting to fall again as the market becomes saturated.  Currently, we are in the lower part of the S curve, but we expect EV sales to eventually make up the vast majority of new light vehicle registrations (cars and vans), which could happen early next decade in some scenarios, later in others.

A limiting factor is the historical rate at which the fleet turns over, sitting around 5% per annum, so even if all new registrations are electric, it would still take over two decades for the NZ fleet to become 100% electric. I can foresee a day, however, when owners of petrol and diesel-fuelled vehicles look around and think to themselves: “if I don’t sell my car soon, and buy an EV, my gas-guzzler’s going to be worth next to nothing.”

One thing I notice is that the 50 kW EV chargers which are still called (somewhat euphemistically) ‘fast’ chargers, are no longer looking so fast. Our two EVs both charge at well over 100 kW (closer to 200 kW actually) and some brands available in NZ are capable of charging at over 500 kW.

When we charge at a 50 kW charger, which being older also tends to be less reliable than the larger ‘hyper’ and ‘super’ chargers, the wait is noticeably longer and one’s mind wanders to thoughts of ‘that 200 kW charger at the… would be better than this’.  It’s probably like the good old days when my great grandparents, who lived in the country, pulled up at the gas station and had to pump the fuel from the bowser by hand, in the knowledge that city cousins were getting those new-fangled bowsers with electric fuel pumps.

To support the growth in EVs, the charging networks in NZ need to keep upgrading their offerings, so that EV owners can charge reliably and quickly wherever they are. I am sure there is some chicken and egg effect going on here, so I expect that as the EV fleet moves up the S curve, the faster EV chargers will be rolled out and the older chargers phased out. Although it removed the direct subsidy on EVs, the current government introduced a scheme to assist the charging networks deploy more chargers, which will help.

Notwithstanding the historical shrinkage in industrial demand, the Tiwai Pt smelter will restart its fourth potline, adding about 400 GWh per annum, and Fonterra and others are moving away from gas and coal to electricity, due either to market pressure to reduce emissions, or more recently to the dramatic run-down in gas production and reserves.

A wild card is the demand that could arise from data centres, the largest of which is proposed for Southland, a project that could add a few thousand GWh per annum of demand on its own, let alone a range of other data centre projects proposed across the country.

The uncertainty now is not so much will demand growth return, but when will it return and how much will it be? When will EV adoption hit max? When will data centres be built and how big will they be?

We don’t see potential for major increases in demand in the ag-fish-forestry sector, but residential demand is now increasing and is likely to continue increasing as the population grows: there is potential for on-going improvements in energy efficiency, but I suspect these have now largely run their course.

Across the five current demand scenarios used in our long-term forecast, adding together the forecasts in each demand sector, the spread by 2030 is from 43,500 GWh per annum to 53,000 GWh, the latter featuring multiple large data centres and the highest rate of industrial electrification.

We don’t put the same probability on all demand scenarios, as we’ll probably end up somewhere in the middle, but one thing we can be sure of: the uncertainty in where demand will go in future is not reducing anytime soon. We can be reasonably sure that demand will grow, but the degree and timing are as uncertain as ever.

The next article in this series will look at generation and how this has changed since the market started in October 1996.

1. Wikipedia
2. Reported on the Energy News website.